False Positive

Logic of Research Methods

Quantitative, social scientific research methods are predicated on the assumption that an objective reality exists and that the job of the researcher is to uncover or understand the state of that reality. In order to do so, scholars collect data. When research is sound, they can be relatively confident that the data accurately represent the true nature of the world. However, social scientists are aware that there are numerous ways that error can be introduced into research and that a data set is unlikely to be a 100% accurate representation of the population of interest. It is the job of the researcher to minimize the likelihood of error. Error can be minimized methodologically (e.g., ensuring that all instruments are both valid and reliable, ensuring that samples are randomly drawn from the population of interest) and/or error can be minimized statistically (e.g., by setting stringent requirements for concluding that findings are real).

Type I (Alpha) Error Rate

Unfortunately, there is no way to know with 100% certainty the true state of the world. The more data that are collected, and the more rigorous the research methods used, the closer researchers are to feeling confident that their findings are an accurate representation of the world. However, social scientists know that it is possible to make mistakes. Therefore, scholars allow for a certain amount of error in their research. In the case of alpha error, for most (though not all) communication research, the acceptable level for committing an alpha error is 5%. This is called the alpha error rate. An alpha error rate of 5% or .05 means that if the null is true (there is no effect), a researcher will falsely reject the null (conclude that there is an effect) no more than 5% of the time. Therefore, the alpha error rate is the maximum percentage of alpha error that the researcher is willing to accept. In other words, one would not be comfortable concluding that an effect was found if the likelihood of committing a Type I error (false positive) is more than 5%.

...